About RefugeeReady East Africa
An AI-assisted pre-submission reviewer that helps NGOs check refugee-response programs against U.S. State Department PRM requirements before they apply.
- Users
- Program managers, proposal writers, and M&E staff at local and international NGOs designing refugee-response programs in Ethiopia, Kenya, and South Sudan, especially smaller local organizations without dedicated grants teams.
- Problem
- PRM requirements are spread across long general guidelines, region-specific NOFOs, and technical standards. Proposals are often weakened by avoidable gaps: missing required indicators, output-only objectives, too few refugees in the target population, or weak protection and coordination plans. Many local NGOs lack the capacity for a rigorous internal review before the deadline.
- Ground truth
- PRM General NGO Guidelines for Overseas Assistance (FY2024), FY2025 PRM Ethiopia/Kenya/South Sudan concept-note NOFO requirements, the Sphere Handbook (2018) minimum standards, and UNHCR country context for Kenya, Ethiopia, and South Sudan.
- Input
- A concept note or program description (upload or paste), plus country, organization type, sectors, and optional budget, duration, refugee share, and NOFO.
- Output
- A transparent readiness score, verdict, and nine-part review (eligibility, scorecard, indicators, Sphere standards, budget, risks, country fit, red flags, top 3 fixes), with evidence quoted from the concept note and a source cited for every judgment. Users can ask follow-up questions, re-review after edits to track improvement, and download the report.
How the score works
The scorecard uses the five published FY2025 PRM concept-note evaluation criteria for Ethiopia, Kenya, and South Sudan: Quality of program idea (20 points), Appropriate identification of beneficiary population, including vulnerable populations (10), Program feasibility/ability to achieve objectives (10), Organization's experience and capacity (5), and Budget (5), for 50 points total. RefugeeReady converts its reviewer ratings into points transparently: Strong = 100%, Adequate = 65%, Weak = 25%, and Cannot assess = 0% of each criterion's maximum. Those rating-to-point conversions are RefugeeReady's internal method, not an official PRM conversion formula. Eligibility is checked separately because PRM screens eligibility before competitive review. Always verify the current NOFO because criteria can change.
| PRM concept-note criterion | Maximum points |
|---|
Safeguards
- Grounded: the AI may only use the built-in source summaries and must say when something isn't covered, rather than inventing requirements, ceilings, or indicator codes.
- Traceable: every judgment cites its source document and quotes the evidence from the concept note it relied on.
- Explainable scoring: the readiness score is computed with a published formula, not produced by the model.
- Privacy: files are read in the browser; users are warned not to include personal data about refugees.
- Human in the loop: the tool supports, not replaces, the organization's judgment and PRM's official review.
Limitations
- Sources are summarized, not the full documents, and PRM priorities and NOFO terms change from year to year.
- AI reviewers can misread documents; the evidence quotes let users check each judgment.
- It does not check budget arithmetic in detail, organizational registration, or SAM.gov status.
Built for the "Becoming a Policy Conductor" Development Lab, O'Neill School of Public and Environmental Affairs, Indiana University.